MAS flags risks of using AI agents in digital-asset trading

SINGAPORE – As artificial intelligence agents move from merely analyzing data to executing transactions, managing risk becomes paramount, according to Chia Der Jiun, Managing Director of the Monetary Authority of Singapore (MAS). Speaking at the "The Future of Digital Finance: Infrastructure, AI and Regulation" conference, Chia highlighted the question of risk management, particularly around liquidity and collateral, as AI agents increasingly act autonomously.

Chia emphasized that while AI agents offer significant benefits in terms of efficiency, accessibility, and innovation, they must be deployed safely. This necessitates controls over authority, permissions, identity, accountability, and governance to ensure AI agents operate within the scope of user instructions, institutional policies, and risk limits, in compliance with regulatory requirements.

The Monetary Authority of Singapore has published the Safeguards for Agentic Finance at Runtime (SAFR) white paper and released an open-source reference implementation of SAFR to support industry experimentation. Chia stressed that the rapid development of technology necessitates a public-private partnership approach to build a comprehensive framework of standards and good practices.

During the conference, discussions centered around the evolving landscape of AI regulations and their potential impact on the future of digital finance.